[Single File] Allow loading T5 encoder in mixed precision #8778
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
sayakpaul
reviewed
Jul 3, 2024
Comment on lines
-558
to
-559
| if torch_dtype is not None: | ||
| pipe.to(dtype=torch_dtype) |
There was a problem hiding this comment.
At the model level by passing in dtype to their respective loading methods
yiyixuxu
approved these changes
Jul 3, 2024
| @@ -1808,4 +1808,17 @@ def create_diffusers_t5_model_from_checkpoint( | |||
|
|
|||
| else: | |||
| model.load_state_dict(diffusers_format_checkpoint) | |||
|
|
|||
There was a problem hiding this comment.
ohh is this related to this?
#8604 (comment)
There was a problem hiding this comment.
Yeah exactly. When you cast the entire model to fp16 the images will be corrupted. You have to use the logic in from_pretrained.
sayakpaul
pushed a commit
that referenced
this pull request
Dec 23, 2024
* update * update * update * update
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
When loading the T5 encoder in SD3 with
from_pretrainedcertain layers are are kept in fp32. Currently this is not the case with single file loading as all modules are cast totorch_dtypewhen loading the pipeline. This is leading to NaNs being produced by the T5 encoder when loading from single file, leading to a black image.This PR
torch_dtype. This is now only handled at the model level.Single file tests for SD3 need to be added. I'll handle that in a follow up PR.
Fixes # (issue)
#8759
Before submitting
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.